Toxicology in Vitro 27 (2013) 2076–2083

Contents lists available at ScienceDirect

Toxicology in Vitro

journal homepage: www.elsevier .com/locate / toxinvi t

A novel o-naphtoquinone inhibits N-cadherin expression and blocksmelanoma cell invasion via AKT signaling

0887-2333/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.tiv.2013.07.011

⇑ Corresponding author. Address: Universidade Federal do Pará, Rua AugustoCorrea, No. 1 – Guamá, CEP: 66075-110, Belém/PA, Brazil. Tel.: +55 9183147882.

E-mail address: rcm.montenegro@gmail.com (R.C. Montenegro).

Raquel Carvalho Montenegro a,⇑, Marne Carvalho de Vasconcellos b, Gleyce dos Santos Barbosa b,Rommel M.R. Burbano a, Luciana G.S. Souza c, Telma L.G. Lemos c, Letícia V. Costa-Lotufo d,Manoel Odorico de Moraes d

a Instituto de Ciências Biológicas, Universidade Federal do Pará, Rua Augusto Corrêa 01-Guamá, Belém/PA, Brazilb Faculdade de Ciências Farmacêuticas, Universidade Federal do Amazonas, Rua Alexandre Amorim, 330-Aparecida, Manaus/AM, Brazilc Departamento de Química Orgânica e Inorgânica, Universidade Federal do Ceará, Campus do Pici, bloco 940 Bairro Pici, Fortaleza/CE, Brazild Departamento de Fisiologia e Farmacologia, Universidade Federal do Ceará, Av. Cel. Nunes de Melo, 1127 – Rodolfo Teófilo, Fortaleza/CE, Brazil

a r t i c l e i n f o a b s t r a c t

Article history:Received 21 January 2013Accepted 23 July 2013Available online 1 August 2013

Keywords:MelanomaBiflorinInvasion inhibitionN-cadherinAKT-1

The down-regulation or loss of epithelial markers is often accompanied by the up-regulation of mesen-chymal markers. E-cadherin generally suppresses invasiveness, whereas N-cadherin promotes invasionand metastasis in vitro. The aim of this work is to investigate the role of biflorin, a naphthoquinone withproven anticancer properties, on the expression of N-cadherin and AKT proteins in MDA-MB-435 invasivemelanoma cancer cells after 12 h of exposure to 1, 2.5 and 5 lM biflorin. Biflorin inhibited MDA-MB-435invasion in a dose-dependent manner (p < 0.01). Likewise, biflorin down-regulated N-cadherin and AKT-1expression in a dose-dependent manner. Biflorin did not inhibit the adhesion of MDA-MB-435 cells to anytested substrates. Additionally, biflorin blocked the invasiveness of cells by down-regulating N-cadherin,most likely via AKT-1 signaling. As such, biflorin may be a novel anticancer agent and a new prototype fordrug design.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Most cancer deaths are due to the development of metastasesand/or the failure of therapeutic regimens (Zhou et al., 2008; Chuet al., 2007). Melanoma is the most invasive and deadly form ofskin cancer. Patients with advanced melanoma with disseminationto distant organs have very poor prognosis, with a median survivaltime of only 6–9 months and a 3-year survival rate of only 10–15%(Eggermont and Robert, 2012; Balch et al., 2009; Sharma et al.,2009).

Most anticancer drugs are generally aimed to inhibit cellgrowth to slow the growth of the primary tumor or to reduce theexisting tumor burden. For several decades, no substantial progresshas been made in developing effective drugs for treating patientswith advanced-stage melanoma (Atallah and Faherty, 2005; Pérezand Danishefsky, 2007). Recent insights into melanoma biology hasresulted in effective immunotherapy and targeted therapy, such asipilimumab, an anti-CTLA-4 monoclonal antibody, and vemurafe-nib, a BRAF inhibitor, which are changing the treatment paradigmfor metastatic melanoma (Graziani et al., 2012; Chapman et al.,

2011). However, as observed in patients with other tumors, pa-tients undergoing immunotherapy and/or targeted therapy usuallydevelop resistance after a period of time. Thus, the approach totreat any type of cancer should be to target a biological network,not just a single molecule (Shuptrine et al., 2012). As a direct resultof the lack of effective therapeutics, the prognosis for patients withmetastatic disease remains very poor. Thus, the use of agents thatinhibit metastasis could be effective, in combination with currentdrugs, to prevent the migration, invasion or colonization of the pri-mary tumor cells at other sites of the body.

To invade, cancer cells of epithelial origin have to migrate fromthe primary tumor mass by breaking their cell–cell contacts,known as adherens junctions (Hazan et al., 2004; Makrilia et al.,2009). The cell adhesion molecule, E-cadherin, a cell-surface pro-tein that accounts for cell-to-cell or cell-to-extracellular matrix(ECM) interactions, is usually absent or dysfunctional in most ofthe advanced, undifferentiated and aggressive breast and otherepithelial carcinomas. This is usually associated with poor patientprognosis (Gupta et al., 2006). Moreover, the loss of E-cadherin intumor cells confers an invasive or metastatic phenotype (Onderet al., 2008). In tumor cells, the gain of expression of another adhe-sion molecule, N-cadherin, has been associated with increasedinvasive potential (Nieman et al., 1999; Hulit et al., 2007; Hazanet al., 2000; Rieger-Christ et al., 2004). Previous studies have

Fig. 1. Chemical structure of biflorin (6,9-dimethyl-3-(4-methyl-3- pentenyl)naph-tha[1,8-bc]-pyran-7,8-dione).

R.C. Montenegro et al. / Toxicology in Vitro 27 (2013) 2076–2083 2077

shown that N-cadherin is up regulated in more invasive breast can-cer cell lines that lacks E-cadherin (Nieman et al., 1999) and that itelicits bladder cell invasion in vitro (Rieger-Christ et al., 2004).

The down-regulation or loss of epithelial markers, such as E-cadherin, is accompanied by the up-regulation of mesenchymalmarkers, such as N-cadherin and vimentin (Yang et al., 2006). Thisprocess is called the epithelial to mesenchymal transition (EMT)and is known to enhance cell motility. As such, E-cadherin gener-ally suppresses invasiveness, whereas N-cadherin promotes inva-sion and metastasis in vitro (Nieman et al., 1999; Hazan et al.,2000).

Multiple factors can induce and regulate the motility oftumorcells, there by contributing to invasion. It has been suggested thatthe expression of N-cadherin is sufficient to trigger EMT, at least bythe activation of PI3-K/Akt pathway (Rieger-Christ et al., 2004). TheAkt family of kinases, i.e., Akt1, Akt2, and Akt3, plays arolein pro-cesses that are well known as hallmarks of cancer, such as sus-tained angiogenesis, unlimited replicative potential, and tissueinvasion and metastasis (Hanahan and Weinberg, 2011). Moreover,Akt activation mediates the expression of N-cadherin and metallo-proteinases and plays aroleintum or invasion and metastasis byinducing EMT (Park et al., 2001; Higuchi et al., 2001; Grille et al.,2003; Wallerand et al., 2010). Recently, Steelman et al. (2011)demonstrated that the activation of AKT-1 increased the resistanceof MCF-7 cells to radiation. Additionally, Toker and Yoeli-Lerner(2006) showed that Akt1 might have a dual role in tumorigenesis,not only promoting it by suppressing apoptosis but also inhibitingit by suppressing invasion and metastasis.

The specific role of AKT in terms of cell motility and invasionseems to depend on the cell type and the pathways that are acti-vated. Many of the enzymes that either mediate the Akt signal,such as MDM2 (Zhou et al., 2001), or regulate Akt activity, suchas the tumor suppressor PTEN (Li et al., 1997), are frequently mu-tated in human tumors. As such, Akt activity is up-regulated, thusincreasing tumor cell growth and survival. In several mammaliansystems, activated Akt1 correlates with cell migration and inva-sion. While constitutively active Akt1 can enhance the ability ofsome cells to invade (Steelman et al., 2011; Kim et al., 2001; Arbol-eda et al., 2003), Akt1 can also have the opposite effect in normal orless invasive cells (Arboleda et al., 2003). Moreover, the increasedactivation of Akt1 correlates with increased proliferation andanchorage-independent growth. However, the effects of activatedAkt1 on cell migration and invasiveness depend on the type of cellsand tissues in which its action is being studied (Steelman et al.,2011; Kim et al., 2001; Arboleda et al., 2003; Enomoto et al.,2005; Irie et al., 2005; Yoeli-Lerner et al., 2005). Yoeli-Lerneret al. (2005) and Toker and Yoeli-Lerner (2006) revealed that theexpression of activated Akt1 potently blocks the migration andinvasion of three distinct breast cancer cell lines through Matrigelin vitro. In fibroblasts, Akt signaling enhances the activation of var-ious small GTPases, leading to remodeling of the actin cytoskeletonand enhancing cell motility (Enomoto et al., 2005). Similarly, theexpression of activated Akt in fibrosarcoma or pancreatic cancercells increases their ability to invade through Matrigel (Parket al., 2001; Kim et al., 2001). Liu et al. (2006) demonstrated thatcells expressing activated Akt1 show increased proliferation andresistance to apoptosis. Additionally, the invasiveness and motilityof the cells were substantially decreased by the down-regulation ofRho activity. Moreover, AKT1 deletion markedly delayed murinebreast tumor formation driven by activated RAS (Skeen et al.,2006). Although the expression of activated AKT1 acceleratesHER-2/NEU-driven breast tumor formation, the tumors that devel-oped were highly differentiated, poorly invasive, and rarely metas-tasized (Hutchinson et al., 2004). AKT1 also plays a prominent rolein tumor angiogenesis. Normal endothelial cells with sustainedactivation of AKT1 develop the complex structural and functional

abnormalities that are characteristic of tumor blood vessels (Phunget al., 2006). These results reinforce the idea that an understandingof cell- and tissue-specific signaling pathways is critical for evalu-ating the implications of activated upstream signaling moleculeson complex phenotypic effects.

Until now, despite large research efforts in targeting tumormetastasis, no progress has been achieved in efficiently preventingmetastasis (Christofori, 2006). This might be due to the mecha-nisms involved in cell migration, which can be reprogrammed,thus allowing the cells to maintain their invasive properties viamorphological and functional de-differentiation.

Natural products have been remarkable source for new antican-cer drugs. Biflorin (Fig. 1), an ortho–naphthoquinone, can be iso-lated from the roots of Capraria biflora L. (Schrophulariaceae),aperennial shrub that was originally found in the Antilles and SouthAmerica (Acosta et al., 2003). This quinone has been shown to haveanticancer properties in vitro and in vivo, increasing the survival ofmice with melanoma tumors, without diminishing the tumor size(Vasconcellos et al., 2005, 2007, 2011). As such, the aim of thiswork is to investigate the role of biflorin in MDA-MB-435, an inva-sive melanoma cancer cell, in vitro.

2. Material and methods

2.1. Cell lines, antibodies, and reagents

The MDA-MB-435 (human melanoma), MCF-10A (normal hu-man breast) and melan-A (normal mouse melanocyties) cell lineswere obtained from American Type Culture Collection (ATCC). Allthe cell lines were cultured according to ATCC recommendations.The following reagents were used: mouse monoclonal anti-N-cad-herin AB-2 (Cell Signaling), mouse monoclonal anti-b-actin (CellSignaling), and Super Signal West Pico Chemiluminescent kit(Pierce). Etoposide, dimethyl sulfoxide, paraformaldehyde, andcrystal violet were purchased from Sigma–Aldrich. Alamar Bluewas purchase from Invitrogen. The invasion plates were obtainedfrom Corning.

2.2. In vitro studies

Cell viability assays.

2.2.1. Alamar Blue assayMDA-MB-435 cells were seeded in 96-well plates at a density of

104 cells per well. They were treated with biflorin, and the AlamarBlueTM assay was performed (Ahmed et al., 1994) after 8, 12, and24 h. After the cells were allowed to attach for 24 h, biflorin (0.1,0.5, 1, 5 and 10 lM) was dissolved in dimethyl sulfoxide (DMSO)and added to each well, and the cells were incubated for 8, 12and 24 h. Etoposide (10, 20 and 50 lM) was used as positive con-trol. Control groups received the same amount of DMSO (0.1%).One hour before the end of the incubations, 20 lL of Alamar Blue™

2078 R.C. Montenegro et al. / Toxicology in Vitro 27 (2013) 2076–2083

was added to each well. The fluorescent signal was monitoredusing a multiplate reader using an excitation wavelength of 530–560 nm and an emission wavelength of 590 nm. The fluorescentsignal generated from the assay was taken to be proportional tothe number of living cells in the sample, as stated by themanufacturer.

2.2.2. Trypan blue exclusionCell viability and death was determined by the trypan blue as-

say (Louis and Siegel, 2011). Cells were seeded in a 12-well plate ata density of 2 � 105 cells per well. After 24 h, they were treatedwith biflorin at 1, 2.5 and 5 lM. Aliquots from each well were re-moved from the cultures after 8, 12, 24 h of incubation, stainedwith 0.4% trypan blue and counted with the Countess™ automatedcell counting platform from Invitrogen. The staining was used toquantify the number of living cells in the samples.

2.2.3. Crystal violet stainCells were seeded in 12-well plates at a density of 2 � 105 cells

per well and treated with biflorin at 1, 2.5 and 5 lM. After 12 h ofincubation, the cells were washed with phosphate buffered saline(PBS), and fixed in 4% paraformaldehyde for 30 min at 4 �C. Thecells were then washed three times with distilled water, and0.1% crystal violet was added to each well. They were then incu-bated for 20 min at room temperature. The plates were washedwith distilled water to remove excess dye and then dried at roomtemperature. The plates were scanned and the intensity of thestained wells was obtained.

2.3. Cell-substrate adhesion assays

For the cell adhesion assay, 96-well plates (Nunc, Roskilde)were coated with Fibronectin, type I and IV collagen by incubatingthe dishes overnight at 4 �C. Any uncoated surfaces of the disheswere blocked by the addition of 2% bovine serum albumin (BSA)(RIA grade; Sigma), which was also used as a negative control.The unbound ECM substrates were removed and the coated disheswere blocked with BSA for 1 h at 37 �C. Then, the dishes werewashed with PBS and media was added before the cells were pla-ted. The MDA-MB 435 cells treated with 1, 2.5 and 5 lM biflorinwere trypsinized, and 4 � 105 cells were plated into each well.After incubation at 37 �C for 2 h, the nonattached cells were re-moved and the remaining cells that were attached were fixed withPHA, washed, and stained with crystal violet. The absorbance wasmeasured at 570 nm. Each panel is representative of duplicateexperiments conductedin triplicate.

2.4. Invasion assay

In vitro invasion assays were performed using modified Boydenchambers consisting of transwell membrane filter inserts (8 lmpore size; Corning Costar Corp., Cambridge, MA, USA) placed in24-well tissue culture plates. The upper surfaces of the membraneswere coated with Matrigel and placed into 24-well tissue cultureplates containing 600 lL of conditioned DMEM media (experimen-tal) or non-conditioned DMEM (control). The cells were seeded inp100 plates (2 � 105 cells/mL) and treated with biflorin at 1, 2.5and 5 lM. After 8 h of treatment, the cells were trypsinized,counted and added to each transwell chamber. They were then al-lowed to invade towards the underside of the membrane for 4 h at37 �C. The cells that passed through the membrane were fixed inmethanol, stained with crystal violet and counted under a lightmicroscope. All assays were performed in duplicate.

2.5. Molecular biology studies

2.5.1. Western blottingThe cells were washed twice with ice-cold PBS and lysedon ice

for 30 min in lysis buffer (100 mmol/L sodium orthovanadate,100 mmol/L NaF, 20 mmol/L HEPES (pH 7.5), 150 mmol/L NaCl,1.5 mmol/L MgCl2, 5 mmol/L sodium pyrophosphate, 10% glycerol,0.2% Triton X-100, 5 mmol/LEDTA, 1 mmol/L phenylmethylsulfonylfluoride, 10 lg/mL leupeptin, and 10 lg/mL aprotinin). The lysateswere clarified by centrifugation at 14,000 g for 10 min at 4 �C.Equal amounts of protein were subjected to SDS–PAGE and trans-ferred tonitrocellulose membranes (Amersham Pharmacia Bio-tech). The membranes were blocked with 5% nonfat milk in TBS-Tween 20 for 1 h at room temperature and then probed with pri-mary antibodies overnight at 4 �C. After incubation with horseradish peroxidase–conjugated secondary antibodies, the immuno-reactive bands were visualized using the Super Signal West PicoChemiluminescent kit (Pierce). Three independent experimentswere performed to analyze the protein levels.

2.5.2. Real-time PCRTotal RNA was extracted from MDA-MB-435 cells with the

RNeasy Mini Kit (Qiagen).Single-stranded cDNA was constructedusing Superscript III polymerase (Invitrogen) and oligo-dT primers.Real-time polymerase chain reaction (RT-PCR) was performed usingthe MyIQ (Bio-Rad) and SYBR Green PCR master-mix reagents (USB).The following primers were used: AKT-1 forward 50-ATGGCACCTT-CATTGGCTAC-30 and reverse 50-AAGGTGCGTTCGATGACAGT-30.

2.6. Statistical analysis

The data are presented as the mean ± SEM. The differences be-tween the experimental groups were compared by analysis of var-iance (ANOVA), followed by Dunnett’s Multiple Comparison Test(p < 0.05) using the GRAPHPAD program (Intuitive Software for Sci-ence, San Diego, CA, USA).

3. Results

3.1. Biflorin was not cytotoxic towards MDA-MB-435 cancer cells andMCF10-A and Melan-A normal cell lines in vitro

MDA-MB-435, an invasive melanoma cancer cell line, was treatedwith increasing concentrations of biflorin, 5, 10 and 20 lM, for 24, 48and 72 h and analyzed by the Alamar Blue™ assay. A significant sup-pression of cell growth was observed in the presence of biflorin(Fig. 2A). To determine the concentration and time required for biflo-rin to inhibit the invasion of MDA-MB-435 cells in the Matrigel modelwithout killing the cells, decreased concentrations of biflorin, 0.1, 0.5,1.0 and 5 lM, were tested for 8, 12 and 24 h and analyzed using theAlamar Blue™assay (Fig. 2B). No cytotoxicity was observed for alltested concentrations at 8 and 12 h. For both experiments, 10, 20and 50 lM etoposide were used as positive controls.

All subsequent experiments were performed in MDA-MB-435cancer cells after 12 h of incubation with 1, 2.5 and 5 lM biflorin.To access cell viability, two models were used, direct cell countingby trypan blue exclusion and colony staining by crystal violet dye.The results indicated that biflorin did not alter the viability of allthree cell lines in all tested concentrations after 12 h of treatment(p > 0.05) (Fig. 3A–F).

3.2. Biflorin inhibits the invasion, but not the adhesion, of MDA-MB435melanoma cancer cells in vitro

Cell invasion is one of the steps involved in metastasis. To deter-mine whether biflorin was involved in this process, the authors

Fig. 2. Biflorin inhibit MDA-MB-435 melanoma cell growth. (a) Biflorin inhibits cellgrowth at all tested concentration and time course (p < 0.05). Etoposide used as apositive control inhibit cell growth at all tested concentration, however only after48 and 72 h (�p < 0.05). Mean values of three independent measurements (±SD) areshown. (b) Biflorin did not inhibit cell growth after 8 and 12 h at all testedconcentration. Only after 24 h, biflorin inhibit cell growth at 5 lM. Etoposide didnot inhibit cell growth. Mean values of three independent measurements (±SD) areshown. ANOVA followed by Tukey’s test.

R.C. Montenegro et al. / Toxicology in Vitro 27 (2013) 2076–2083 2079

first ensured that the inhibition of invasion was not due tocell death. Thus, the viability of MDA-MB-435 melanoma cancercells was assessed after 8 and 12 h of treatment with 1, 2.5 and5 lM biflorin. As shown in Fig. 3C, cell death was not observed inany of the concentrations of biflorin and durations of incubationtested. However, a strong and dose-dependent reduction in theinvasion of MDA-MB-435 cells through the Matrigel matrix wasobserved after the treatment with 1, 2.5 and 5 lM biflorin(38.25 ± 9.53; 16.5 ± 3.31 and 2.25 ± 0.95, respectively).In compar-ison, this was not observed in the negative control (55.00 ± 3.9)(Fig. 4A and B). Additionally, biflorin did not inhibit the adhesionof MDA-MB-435 cells to any of the ECM substrates tested (datanot shown).

3.3. Biflorin inhibits invasion by down-regulating N-cadherin and AKT-1

The cadherins are a family of a cell to cell adhesion moleculesthat have been implicated in the invasive process (Hanahan andWeinberg, 2011). To determine whether the inhibition of invasionby biflorin was related to N-cadherin protein levels, a western blotwas performed. After 12 h of biflorin treatment, the protein levelsof N-cadherin were down-regulated in a dose dependent manner(Fig. 4C and D). To further understand the signaling pathways in-volved in the inhibition of invasion, the expression levels of AKT-1 was assessed. 36B4, acidic ribosomal phosphoprotein P0, wasused as a reference gene. AKT-1 mRNA levels were down-regulatedin a dose-dependent manner by 94.65, 76.25 and 21.35%, by 1, 2.5and 5 lM biflorin, respectively (Fig. 4E). After 12 h of treatment

with 5 lM biflorin, the AKT-1 (p < 0.05) mRNA level was decreasedby 5-fold (p < 0.05).

4. Discussion

Melanoma is one of the most invasive and deadly forms of skincancer, and only a few agents are available for treating advanceddisease to enable long-term patient survival. However, theseagents are relatively ineffective, with overall response rates of 5–20%. This finding supports the need for identifying new compoundsthat inhibit the pathways that are deregulated in melanoma(Eggermont and Robert, 2012; Sharma et al., 2009).

Anticancer drug development strategies are usually aimed at di-rectly inhibiting the growth of the primary tumor or reducing theexisting tumor burden. Therapeutic agents that can inhibit metas-tasis could be an option for preventing colonization, thereby en-abling the containment of the primary tumors in a chemicallymanageable form (Pérez and Danishefsky, 2007; Hedley et al.,2004).

In this study, using melanoma cell lines as a model for invasionstudies, we investigated the ability of biflorin, an ortho-naphtho-quinone, to treat solid tumors. We also investigated the EMC sub-strates, Fibronectin and types I and IV collagen, and the expressionof N-cadherin and AKT1.

The MDA-MB-435 cell line is a controversial invasive cancer cellline because, for several decades, it was considered to be a breastcancer cell line (Chambers, 2009). However, gene expression profil-ing showed that it expressed both epithelial and melanocyticmarkers, and as such, the National Cancer Institute (NCI) now con-siders it to be a melanoma cell line (Garraway et al., 2005). Regard-less of its origin, the MDA-MB-435 cell line is an aggressive cell linethat can be used as a model to study invasion in vitro (Hulit et al.,2007).

Biflorin has been shown to be cytotoxic to cancer cell linesin vitro and in vivo (Vasconcellos et al., 2005, 2007, 2011, 2010). Re-cently, Vasconcellos et al. (2011) described that biflorin inhibitsseveral melanoma cell lines in vitro. Moreover, biflorin inhibitedDNA synthesis, leading to the apoptosis of B16, a murine mela-noma cell line (Vasconcellos et al., 2011). Biflorin also increasedsurvival rates and inhibited tumor growth in an in vivo melanomaB16 model (Vasconcellos et al., 2011). These results supportedgaining an understanding of the mechanisms behind biflorin activ-ity in cancer cells.

Here, we described that biflorin inhibits MDA-MB-435 cell inva-sion in vitro in a dose-dependent manner. The MDA-MB-435 cellline expresses genes associated with both breast cancer and mela-noma. For this reason, the authors conducted the experiments onnormal cell lines of both breast and melanocyteorigin, MCF-10Aand Melan-A, respectively. However, at all concentrations tested,biflorin had no effect on the normal cell lines. Vasconcellos et al.(2010) demonstrated that at lower concentrations, biflorin hadsig-nificant antioxidant and protective effects against cytotoxicity,genotoxicity, mutagenicity, and intracellular lipid peroxidation in-duced by H2O2 in yeast and normal mammalian cells. This resultcould be attributed to its hydroxyl radical-scavenging propertyand corroborated our findings. However, at higher concentrations,biflorin was cytotoxic and genotoxic.

Many changes in gene expression and protein functions occurduring tumor progression. Alterations in cell–cell and cell-matrixadhesion seem to have a central role in facilitating the migration,invasion and metastatic dissemination of tumor cells (Yilmaz andChristofori, 2010). Multiple cell adhesion molecules, such asE- and N-cadherin, which are calcium-dependent cells adhesionmolecules that mediate homophilic cell–cell adhesion, have beenreported to play roles in melanoma progression. The cadherin

Fig. 3. Biflorin did not inhibit cell growth at 12 h treatment. (a and b) Results of MCF-10A cell viability. (c and d) Results of Melan-A cell viability. (e and f) Results of MDA-MB-435 cell viability; all evaluated by the trypan blue and crystal violet, respectively. Three independent experiments were performed in triplicate (mean ± SE). No significantdifferences were observed for all cells treated with biflorin at 12 h. ANOVA followed by Tukey’s test.

2080 R.C. Montenegro et al. / Toxicology in Vitro 27 (2013) 2076–2083

switch from E-cadherin to N-cadherin results in the disassociationof melanoma cells and promotes the invasion of melanoma cells(Hsu et al., 2000). The pro-invasive action of N-cadherin persistedeven in the presence of E-cadherin, suggesting that N-cadherin hasa dominant effect over the normally tumor-suppressive E-cad-herin, (Nieman et al., 1999; Hazan et al., 2000). Additionally, N-cadherin plays a role in the formation and stability of blood vessels(Paik et al., 2004), which are processes that are important for tu-mor growth (Naumov et al., 2008).

It is well known that N-cadherin is upregulated in invasive tu-mor cell lines and in tissues from melanomas, breast and prostatecancers (Hazan et al., 1997).Inhibitors of N-cadherin function havebeen demonstrated to cause apoptosis in certain cell types (Erezet al., 2004), and drugs targeting N-cadherin may thus have multi-ple therapeutic applications.

The MDA-MB-435 cell line over expresses N-cadherin, but itdoes not express E-cadherin. As such, it is a suitable model for

studying processes related to cell motility, invasion and metastasis(Nieman et al., 1999). In our model, biflorin decreased the expres-sion of N-cadherin in a dose-dependent manner, thus accountingfor its inhibitory effect on invasion. Our results are supported bythose of Lee et al. (1998). As such, we propose that the effect ofbiflorin on the invasiveness of this cell line is most likely due toits action on the expression of N-cadherin. These results may ex-plain the increased survival in mice treated with biflorin thatwas observed by Vasconcellos et al. (2011). Given the results ob-tained so far, we propose a mechanism underlying biflorin action(Fig. 5).

N-cadherin antagonists have shown promise as anti-canceragents in pre-clinical and clinical trials (Lyon et al., 2010; Beasleyet al., 2009). One of the major issues to be resolved is why N-cad-herin antagonists, such as ADH-1 and anti-N-cadherin Mabs, arenot toxic, given the wide distribution of this cell adhesion molecule(CAM) (Blaschuk, 2012). This is also the case for biflorin because no

Fig. 4. Biflorin interaction impacts on invasive activity of melanoma cells via modulating N-cadherin and AKT-1 expression in vitro. (a) Representative images of theTranswell assays of MDA-MB-435 cell invasion. Magnification 40� (1-control, 2–1.0 lM, 3–2.5 lM, 4–5.0 lM). (b) Results of Transwell assays showed the relative number ofinvasive MDA-MB-435 cells with indicated treatments after 12 h. Three independent experiments were performed in triplicate (mean ± SE). �p < 0.01. (c) Expression of N-cadherin in melanoma cells was detected by Western blotting after 12 h. Three independent experiment were performed. (d) N-cadherin protein densitometry. Fourindependent experiment were performed. �p < 0.05. ANOVA follow by Tukey’s. (e) Real time RT-PCR results of mRNA expression of AKT-1 in melanoma cells treated withbiflorin after 12 h. Three independent experiments were performed in triplicate (mean ± SE). �p < 0.05. ANOVA followed by Tukey’s test.

R.C. Montenegro et al. / Toxicology in Vitro 27 (2013) 2076–2083 2081

toxicity to normal cells were observed, making it a promising CAMinhibitor or drug lead.

It has been suggested that the expression of N-cadherin is suf-ficient to trigger EMT, at least in part, by the activation of thePI3-K/Akt pathway (Rieger-Christ et al., 2001). Moreover, N-cad-herin and phospho-EGFR expression have been associated withAkt activation and with the modulation of invasiveness (Wallerandet al., 2010).

AKT is a serine-threonine kinase whose isoforms, AKT1, AKT2,and AKT3, exist in mammalian cells. These play roles in processesthat are considered hallmarks of cancer, such as unlimited replica-tive potential, sustained angiogenesis, tissue invasion and metasta-sis (Hanahan and Weinberg, 2011).

A functional role for Akt in cell motility was first reported byMeili et al. (1999). Servant et al. (2000) subsequently demon-strated similar effects in neutrophils. Although both AKT1 and

AKT2 have a role in cell motility and invasion, distinct and evenopposing functions have been described for these proteins. In somecell systems, AKT1 enhances cell invasion and migration. In others,AKT1 inhibits motility, while AKT2 promotes motility (Vasko et al.,2004). Moreover, Akt1 nullfibroblasts have been shown to be lessmotile and invasive when compared to control fibroblasts (Irieet al., 2005; Yoeli-Lerner et al., 2005; Vasko et al., 2004). BothAKT1 and AKT2 were found to enhance the invasiveness of humanpancreatic cancer cells, raising the possibility that the effects ofAkt1 on invasiveness and motility are cell type specific (Skeenet al., 2006). In vivo studies in which tumor formation is reducedby crossing mice into an Akt1-deficient background support theseobservations (Saji et al., 2005). Our results suggest that the inhibi-tion of invasion of the MDAMB-435 melanoma cell line by biflorinis due to the down-regulation of N-cadherin, which inhibits AKT1expression. This observation is in agreement with several other

Fig. 5. The down-regulation or loss of epithelial markers is accompanied bymesenchymal markers neoexpression. E-cadherin generally suppresses invasive-ness, whereas N-cadherin promotes invasion and metastasis in vitro. Biflorin is anaphthoquinone with proven anticancer properties. In this way, the aim of thiswork is to investigate the role of biflorin in MDA-MB-435 invasive melanomacancer cell and the N-cadherin protein status. Biflorin inhibited MDA-MB-435invasion in a dose- dependent manner (p < 0.01). Likewise, biflorin down-regulatedN-cadherin and AKT-1 expression in a dose-dependent manner. Biflorin did notinhibit MDA-MB-435 cell adhesion to all tested substrate. Biflorin can be aprototype to new anticancer drugs design.

2082 R.C. Montenegro et al. / Toxicology in Vitro 27 (2013) 2076–2083

studies (Steelman et al., 2011; Vasko et al., 2004; Saji et al., 2005)that have reported that AKT1 promotes motility in different celllines. Thus, its inhibition could abolish cell invasion, as was ob-served in our model. These observations are particularly important,given the development of both generalized and isoform-specificAkt inhibitors for clinical trials.

In summary, to our knowledge, this is the first mechanisticexplanation for how biflorin, a natural compound, abrogates inva-sion. We showed that in MDA-MB-435 melanoma cells, this mostlikely occurs by the inhibition of N-cadherin and the AKT-1 path-way. Further animal and iRNA studies have to be performed to fullyelucidate the mechanisms underlying the biological actions ofbiflorin.

Conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgements

The authors are grateful to the Brazilian Agencies FINEP, CNPq,CAPES and FAPEAM for fellowships and financial support.

References

Acosta, S., Muro, L., Sacerio, A., 2003. Anti-inflammatory effects of an aqueousextract of Capraria biflora L. Acta. Farm. Bonaerense. 22, 53–55.

Ahmed, S., Gogal Jr., R., Walsh, J., 1994. A new rapid and simple non-radioactiveassay to monitor and determine the proliferation of lymphocytes: an alternativeto [3H] thymidine incorporation assay. J. Immunol. Methods 15, 211–224.

Arboleda, M.J., Lyons, J.F., Kabbinavar, F.F., et al., 2003. Cancer res. 63, 196–206.Atallah, E., Faherty, L., 2005. Treatment of metastatic malignant melanoma. Curr.

Treat Opt. Oncol. 6, 185–193.Balch, C., Gershenwald, J., Soong, S., et al., 2009. Final version of 2009 AJCC

melanoma staging and classification. J. Clin. Oncol. 27, 6199–6206.Beasley, G., McMahon, N., Sanders, G., et al., 2009. A Phase 1 study of systemic ADH-

1 in combination with melphalan via isolated limb infusion in patients withlocally advanced in-transit malignant melanoma. Cancer 115, 4766–4774.

Blaschuk, O., 2012. Discovery and development of N-cadherin antagonists. CellTissue Res. 348, 309–313.

Chambers, A., 2009. MDA-MB-435 and M14 cell lines: identical but not M14melanoma? Cancer Res. 69, 5292–5293.

Chapman, P., Hauschild, A., Robert, C., 2011. Improved survival with vemurafenib inmelanoma with BRAF V600E mutation. N. Engl. J. Med. 364, 2507–2516.

Christofori, G., 2006. New signals from the invasive front. Nature 441, 444–450.Chu, S., Yang, S., Liu, S., et al., 2007. In vitro and in vivo antimetastatic effects of

Terminalia catappa L. leaves on lung cancer cells. Food Chem. Toxicol. 45, 194–1201.

Eggermont, A., Robert, C., 2012. Melanoma in 2011: a new paradigm tumor for drugdevelopment. Nat. Rev. Clin. Oncol. 10, 74–76.

Enomoto, A., Murakami, H., Asai, N., et al., 2005. Akt/PKB regulates actinorganization and cell motility via Girdin/APE. Dev Cell 9, 389–402.

Erez, N., Zamir, E., Gour, B., et al., 2004. Induction of apoptosis in culturedendothelial cells by a cadherin antagonist peptide: involvement of fibroblastgrowth factor receptor-mediated signaling. Exp. Cell Res. 294, 366–378.

Garraway, L., Widlund, H., Rubin, M., et al., 2005. Integrative genomic analysesidentify MITF as a lineage survival oncogene amplified in malignant melanoma.Nature 436, 117–122.

Graziani, G., Tentori, L., Navarra, P., 2012. Ipilimumab: a novel immunostimulatorymonoclonal antibody for the treatment of cancer. Pharmacol. Res. 65, 9–22.

Grille, S., Bellacosa, A., Upson, J., 2003. The protein kinase Akt induces epithelialmesenchymal transition and promotes enhanced motility and invasiveness ofsquamous cell carcinoma lines. Cancer Res. 63, 2172–2178.

Gupta, M., Gupta, S., Hoffman, B., Liebermann, D., 2006. Gadd45a and Gadd45bprotect hematopoietic cells from UV-induced apoptosis via distinct signalingpathways, including p38 activation and JNK inhibition. J. Biol. Chem. 281,17552–17558.

Hanahan, D., Weinberg, R., 2011. The hallmarks of cancer: the next generation. Cell144, 646–674.

Hazan, R., Kang, L., Whooley, B., Borgen, P., 1997. N-cadherin promotes adhesionbetween invasive breast cancer cells and the stroma. Cell Adhes. Commun. 4,399–411.

Hazan, R., Phillips, G., Qiao, R., et al., 2000. Exogenous expression of N-cadherin inbreast cancer cells induces cell migration, invasion, and metastasis. J. Cell Biol.148, 779–790.

Hazan, R., Qiao, R., Keren, R., 2004. Cadherin switch in tumor progression. Ann. N. Y.Acad. Sci. 1014, 155–163.

Hedley, B., Winquist, E., Chambers, A., 2004. Therapeutic targets for antimetastatictherapy. Exp. Opin. Ther. Targets 8, 527–536.

Higuchi, M., Masuyama, N., Fukui, Y., 2001. Akt mediates Rac/Cdc42-regulated cellmotility in growth factor-stimulated cells and in invasive PTEN knockout cells.Curr. Biol. 11, 1958–1962.

Hsu, M., Andl, T., Li, G., et al., 2000. Cadherin repertoire determines partner-specificgap junctional communication during melanoma progression. J. Cell Sci. 113,1535–1542.

Hulit, J., Suyama, K., Chung, S., et al., 2007. N-cadherin signaling potentiatesmammary tumor metastasis via enhanced extracellular signal-regulated kinaseactivation. Cancer Res. 67, 3106–3116.

Hutchinson, J.N., Jin, J., et al., 2004. Activation of Akt-1 (PKB-a) can accelerate ErbB-2-mediated mammary tumorigenesis but suppresses tumor invasion. CancerRes. 64, 3171–3178.

Irie, H., Pearline, R., Grueneberg, D., 2005. Distinct roles of Akt1 and Akt2 inregulating cell migration and epithelial–mesenchymal transition. J. Cell Biol.171, 1023–1034.

Kim, D., Kim, S., Koh, H., et al., 2001. Akt/PKB promotes cancer cell invasion viaincreased motility and metal-loproteinase production. FASEB J. 15, 1953–1962.

Lee, S., Reimer, C., Campbell, D., et al., 1998. H-cadherin expression inhibits in vitroinvasiveness and tumor formation in vivo. Carcinogenesis 19, 1157–1159.

Li, J., Yen, C., Liaw, D., et al., 1997. PTEN, a putative protein tyrosine phosphatasegene mutated in human brain, breast, and prostate cancer [see comments].Science 275, 1943–1947.

Liu, H., Radisky, D., Nelson, C., et al., 2006. Mechanism of Akt1 inhibition of breastcancer cell invasion reveals a protumorigenic role for TSC2. Proc. Natl. Acad. Sci.U S A. 103, 4134–4139.

Louis, K.S., Siegel, A.C., 2011. Cell viability analysis using trypan blue: manual andautomated methods. Methods Mol. Biol. 740, 7–12.

Lyon, C., Koutsouki, E., Aguilera, C., et al., 2010. Inhibition of N-cadherin retardssmooth muscle cell migration and intimal thickening via induction of apoptosis.J. Vasc. Surg. 52, 1301–1309.

Makrilia, N., Kollias, A., Manolopoulos, L., Syrigos, K., 2009. Cell adhesion molecules:role and clinical significance in cancer. Cancer Invest. 27, 1023–1037.

Meili, R., Ellsworth, C., Lee, S., 1999. Chemo attractant-mediated transient activationand membrane localization of Akt/PKB is required for efficient chemotaxis tocAMP in Dictyostelium. EMBO J. 18, 2092–2105.

Naumov, G., Folkman, J., Straume, O., Akslen, L., 2008. Tumor–vascular interactionsand tumor dormancy. APMIS 116, 569–585.

Nieman, M., Prudoff, R., Johnson, K., Wheelock, M., 1999. N-cadherin promotesmotility in human breast cancercells regardless of their E-cadherin expression.J. Cell Biol. 147, 631–644.

Onder, T., Gupta, P., Mani, S., et al., 2008. Loss of E-cadherin promotes metastasis viamultiple downstream transcriptional pathways. Cancer Res. 15, 3645–3654.

Paik, D., Fu, C., Bhattacharya, J., Tilson, M., 2004. Ongoing angiogenesis in bloodvessels of the abdominal aortic aneurysm. Exp. Mol. Med. 36, 524–533.

Park, B., Zeng, X., Glazer, R., 2001. Akt1 induces extracellular matrix invasion andmatrix metalloproteinase-2 activity in mouse mammary epithelial cells. CancerRes. 61, 7647–7653.

Pérez, L., Danishefsky, S., 2007. Chemistry and biology in search of antimetastaticagents. ACS Chem. Biol. 2, 159–162.

R.C. Montenegro et al. / Toxicology in Vitro 27 (2013) 2076–2083 2083

Phung, T.L., Ziv, K., et al., 2006. Pathological angiogenesis is induced by Akt signalingand inhibited by rapamycin. Cancer Cell 10, 159–170.

Rieger-Christ, K., Cain, J., Braasch, J., 2001. Expression of classic cadherins type I inurothelial neoplastic progression. Hum. Pathol. 32, 18–23.

Rieger-Christ, K., Lee, P., Zagha, R., et al., 2004. Novel expression of N-cadherinelicits in vitro bladder cell invasion via the Akt signaling pathway. Oncogene 23,4745–4753.

Saji, M., Vasko, V., Kada, F., et al., 2005. Akt1 contains a functional leucine-rich nuclear export sequence. Biochem. Biophys. Res. Commun. 332, 167–173.

Servant, G., Weiner, O., Herzmark, P., et al., 2000. Polarization of chemoattractantreceptor signaling during neutrophil chemotaxis. Science 287, 1037–1040.

Sharma, S., Kelly, T., Jones, P., 2009. Epigenetics in cancer. Carcinogenesis 31, 27–36.

Shuptrine, C., Surana, R., Weiner, L., 2012. Monoclonal antibodies for the treatmentof cancer. Semin. Cancer Biol. 22, 3–13.

Skeen, J.E., Bhaskar, P.T., et al., 2006. Akt deficiency impairs normal cell proliferationand suppresses oncogenesis in a p53-independent and mTORC-1-dependentmanner. Cancer Cell 10, 269–280.

Steelman, L., Navolanic, P., Chapel, W., 2011. Involvement of Akt and mTOR inchemotherapeutic- and hormonal-based drug resistance and response toradiation in breast cancer cells. Cell Cycle 10, 3003–3015.

Toker, A., Yoeli-Lerner, M., 2006. Akt signaling and cancer: surviving but not movingon. Cancer Res. 66, 3963–3966.

Vasconcellos, M.C., Montenegro, R.C., Militão, G., et al., 2005. Bioactivity of biflorin, atypical o-naphthoquinone isolated from Capraria biflora L. Z. Naturforsch C. 60,394–398.

Vasconcellos, M.C., Bezerra, D., Fonseca, A., et al., 2007. Antitumor activity ofbiflorin, an o-naphthoquinone isolated from Capraria biflora. Biol. Pharm. Bull.30, 1416–1421.

Vasconcellos, M.C., Moura, D., Rosa, R., et al., 2010. Evaluation of the cytotoxic andantimutagenic effects of biforin, an antitumor 1,4 o-naphthoquinone isolatedfrom Capraria biflora L. Arch. Toxicol. 84, 799–810.

Vasconcellos, M.C., Bezerra, D., Fonseca, A., et al., 2011. The in-vitro and in-vivoinhibitory activity of biflorin in melanoma. Mel. Res. 2, 106–114.

Vasko, V., Saji, M., Hardy, E., et al., 2004. Akt activation and localization correlatewith tumor invasion and oncogene expression in thyroid cancer. J. Med. Genet.41, 161–170.

Wallerand, H., Cai, Y., Wainberg, Z., 2010. Phospho-Akt pathway activation andinhibition depends on N-cadherin or phospho-EGFR expression in invasivehuman bladder cancer cell lines. Urol. Oncol. 28, 180–188.

Yang, J., Mani, S., Weinberg, R., 2006. Exploring a new twist on tumor metastasis.Cancer Res. 66, 4549–4552.

Yilmaz, M., Christofori, G., 2010. Mechanisms of motility in metastasizing cells. Mol.Cancer Res. 8, 629–642.

Yoeli-Lerner, M., Yiu, G., Rabinovitz, I., et al., 2005. Akt blocks breast cancer cellmotility and invasion through the transcription factor NFAT. Mol. Cell. 20, 539–550.

Zhou, B.P., Liao, Y., Xia, W., Zou, Y., Spohn, B., Hung, M.C., 2001. HER-2/neu inducesp53 ubiquitination via Akt-mediated MDM2 phosphorylation. Nat. Cell Biol. 3,973–982.

Zhou, J., Chen, Y., Lang, J., et al., 2008. Salvicine inactivates beta 1 integrin andinhibits adhesion of MDA-MB-435 cells to fibronectin via reactive oxygenspecies signaling. Mol. Cancer Res. 6, 194–204.